AUSTRALIAN ARCHAEOLOGY 2019, VOL. 85, NO. 2, 170–183 https://doi.org/10.1080/03122417.2019.1657320 ARTICLE Insights into the procurement and distribution of fossiliferous chert artefacts across southern Australia from the archival record Ingrid Warda , Marcus M. Key Jr.b , Rosine Rierac, Annie Carsond and Michael O’Learyc aSchool of Social Sciences, University of Western Australia, Crawley, Australia; bDepartment of Earth Sciences, Dickinson College, Carlisle, PA, USA; cDepartment of Earth Sciences, University of Western Australia, Crawley, Australia; dWestern Australian Museum, Welshpool, Australia ABSTRACT ARTICLE HISTORY Following from previous research in Western Australia, this study explores the use of Received 2 February 2019 embedded microfossils—including bryozoan and foraminiferal fossil assemblages—to help Accepted 15 August 2019 identify source and distribution of fossiliferous chert artefacts from South Australian archives. KEYWORDS Artefacts from key archaeological sites include Allen’s Cave, Koonalda Cave, Wilson’s Bluff, Fossiliferous chert; sourcing; Ooldea and Kongarong. Preliminary analyses indicate a possible differentiation of fossiliferous microfossils; mobility; chert types east and west of the Eyre Peninsula, from Otway Basin and Eucla Basin limestones, South Australia respectively. The widespread distribution and trade of fossiliferous chert is supported by ethnographic descriptions of the sources, procurement and trade as recorded by Daisy Bates, Thomas Draper Campbell and Norman Tindale. Further work combining biostratigraphy and lithology with archaeology and ethnography is needed to explore these ideas further. Background of the WA Museum Archives have already revealed examples of fossiliferous chert artefacts from regions Chert, or flint as it is often synonymously termed, much further north and east than initially assumed has been used for making stone tools for thousands ’ of years, being associated with the earliest fossil (Key et al. 2019;OLeary et al. 2017;Wardetal. remains in Africa (Hershkovitz et al. 2018) through 2019). At the same time, preliminary studies of diag- to small, fine artefacts in the Neolithic and later nostic fossil bryozoan assemblages within samples of periods of France and England (Beasley 1984; chert debitage from southwest Western Australia Cunliffe 2001:89–90). It has also been used in more demonstrated that the source could not be offshore recent times for industrial and armament purposes in the Perth Basin as previously purported (Glover (e.g. Flint et al. 1989:20). In Australia, chert artefacts 1975a, 1975b) but onshore in the Eucla Basin ’ are documented in several early sites, most notably (O Leary et al. 2017). However, the idea of long-dis- Devil’s Lair (Dortch 1984) and Allen’s Cave (Cane tance trade of fossiliferous chert with Aboriginal 1995; Marun 1972), and also in sites dating to the groups westward continues to be dismissed due to early to mid-Holocene (Beasley 1984; Cann et al. the apparently large distances involved (Dortch and 1999; Coutts 1981; McNiven et al. 2012). Indeed in Dortch 2019; Glover 1975a). This is despite evidence SW Western Australia, fossil-bearing chert has been of trade of this material up to 700 km eastward used, arguably erroneously (see O’Leary et al. 2017), (Bates 1921;Johnston1941; Nicholson and Cane as a chronological divide between early to mid- 1991; see also Bourman et al. 2016:167; Ward et al. Holocene age sites (Dortch 1986; Dortch and 2019), and also the distribution of lithic raw materi- McArthur 1985; Glover et al. 1993; Hallam 1987; als, edge-ground axes, wooden items, ochre and baler Marwick 2002; Smith 1993). shell hundreds of kilometres over much of Australia More important than the chronological context is (see McBryde 1997;Muntetal.2018:79 the spatial context of chert-bearing archaeological and references). sites and in particular those found inland from Having previously focused on fossiliferous chert coastal sources (Beasley 1984; see also below) as archives from Western Australia, this study explores these imply human dispersal or trade. Investigations the South Australian archival records1 and what this CONTACT Ingrid Ward [email protected] College School of Social Sciences, University of Western Australia, M257 Archaeology, UWA Crawley 6009, Australia 1Attempts were also made to explore the archival record in Victoria. Queries of the Victorian State heritage databases revealed 2889 registered Aboriginal places where chert has been identified across the State. These registers do not discriminate biogenic and non-biogenic chert. Viewing these records is complicated by the need to obtain approval of all registered Aboriginal parties, as well as the approval of the Victorian Aboriginal Heritage Council (VAHC). Hence it was decided just to focus on the South Australian archival collections for this study. ß 2019 Australian Archaeological Association AUSTRALIAN ARCHAEOLOGY 171 Figure 1. Location of sites and basins mentioned in the text. might also reveal about the use and distribution of Terminology this diagnostic material. In particular, we focus on the As outlined by Munt et al. (2018) there is some con- records of Norman Tindale, Daisy Bates and Thomas fusion in the use of the terms ‘chert’ and ‘flint’ for Draper Campbell, each of whom recorded and col- microcrystalline or cryptocrystalline silica (e.g. lected chert artefacts from sites along the Eucla coast Luedtke 1992;Whittaker1994:70; see also Sieveking and in other parts of South Australia between 1930 and Hart 1986). In Koonalda Cave, Marun (1972) and late 1960. As part of this investigation we identify identifies flint and ‘numerically insignificant’ quanti- bryozoan fossils on the surface of artefacts from key archaeological sites, including Allen’s Cave, Koonalda ties of chert and limestone, whilst Cane (1995)and ’ Cave, Wilson’s Bluff, Ooldea and a few other sites in Wright (1971) use only the term flint. For the Allen s South Australia (Figure 1). This follows a similar Cave site, Munt et al. (2018:74) use chert for lighter ‘ ’ study undertaken by Benbow and Nicholson (1992), coloured material and flint for denser, blacker who also undertook a lithological and microscopic material that is more like the flint from the English comparison of artefacts from the South Australian chalks from which the term derived (e.g. Rapp Museum and Moeller collections with silicified source 2009:76, 79). Glover (1975a) similarly preferred the ‘ ’ material from Wilson’s Bluff, Koonalda Cave and term chert over the unsatisfactorily defined flint to Head of the Bight. Whilst distinct lithologies were describe artefacts from SW Western Australia, observed in each of the source materials, and in par- including Devils Lair. In the Lake Condah region of ticular the presence of relict bryozoan clasts in the Victoria however, McNiven et al. (2012:50) differenti- Wilson Bluff Limestone, it was found that the vari- ate (non-fossiliferous) chert from cherts that ‘contain ation in degree of silicification precluded provenanc- microfossils consistent with flint’. ing artefacts to any of these particular sources Both chert and flint are formed by diagenetic (Benbow and Nicholson 1992). We argue that the replacement of carbonate by silica under high pres- identification of the embedded fossils is the key to sure by burial beneath the seafloor. Flint is generally provenancing fossiliferous chert artefacts. We preface reserved for material which has originated from this discussion with a brief consideration of the ter- chalk deposits (or marl) and chert for material from minology and stratigraphic context of chert-bearing limestone deposits, the latter usually formed from limestones along the southern coastline of Australia. the skeletons of organisms with a calcium carbonate 172 I. WARD ET AL. biomineralogical structure. This explains the foraminifera (James and Bone 1991;O’Connell et al. description of ‘silicified limestone’—or more specif- 2012). According to Benbow and Nicholson (1992:3), ically ‘silicified bryozoan limestone’—by Benbow the younger Miocene limestone has been locally silici- and Nicholson (1992) for artefacts from the fied in groundwater environments around the mar- Nullarbor Plain. Similarly, in SW Western Australia, gins of the Nullarbor Plain, such as in the Yarle Glover (1974, 1975b; see also Martin 1982) makes Lakes—Ooldea area. Further east is the Upper Eocene the distinction between ‘fossiliferous cryptocrystal- to Middle Miocene Gambier Limestone within the line chert’ and ‘palaeontologically barren’ chert. Gambier and Otway Basins, which also contains fos- Indeed Martin (1982) suggests there may be up to siliferous chert (Blakemore et al. 2014; Bourman et al. five types of chert in SW Western Australia, differ- 2016). The latter outcrops east of Naracoorte scarp entiated by presence/absence, type and abundance and in coastal areas near Mount Gambier (Sexton of embedded fossil species. Hence the more descrip- 1965). Eocene chert and silica-rich rocks have also tive terminology, rather than simply chert, is been observed in the Bremer Basin and in the St important in terms of characterising the use and Vincent Basin (Li et al. 2000). provenance of these silicified artefacts. Here we fol- Fossiliferous chert occurs as horizontal seams within ’ low Glover s(1975a, 1975b) example and use the these limestone sequences, which are present along term fossiliferous chert to describe archival samples. coastal cliffs (Figure 2) and in some cave profiles.